Mistral on NVIDIA Ampere

Can I run Mistral Small 24B on an RTX 3080 10GB?

Not at Q4_K_M — it needs 15.5 GB against 8.6 GB available. You would need 2 of these cards.

Won't fit

15.5 GB of 8.6 GB · 180%
016 GB
Weights 13.4 GB
KV cache 1.3 GB
Runtime overhead 0.8 GB
Over the limit 6.9 GB

Short by 6.9 GB. You can run it with 19 of 40 layers on the RTX 3080 10GB and the rest in system RAM, at roughly 5.03 tok/s — usable for batch work, painful for chat. A smaller quantisation or a shorter context is usually the better trade.

Generation5.03tok/s
Prompt processing535tok/s
Max context0tokens
KV per 1K tokens0GB

Every quantisation of Mistral Small 24B on a RTX 3080 10GB

Highlighted row is the highest quality that still fits at 8K context.

QuantisationbpwWeights TotalVerdictMax ctxtok/s
FP16 / BF16 16.00 43.9 GB 46.0 GB Won't fit 1.01
INT8 / W8A8 8.50 23.0 GB 25.1 GB Won't fit 2.26
Q8_0 (GGUF) 8.50 23.0 GB 25.1 GB Won't fit 2.26
FP8 (E4M3) 8.00 22.0 GB 24.0 GB Won't fit 2.36
Q6_K 6.56 18.0 GB 20.1 GB Won't fit 3.16
Q5_K_M 5.67 15.6 GB 17.6 GB Won't fit 3.89
Q5_K_S 5.52 15.1 GB 17.2 GB Won't fit 4.14
AWQ 4-bit 4.25 13.5 GB 15.6 GB Won't fit 5.00
GPTQ 4-bit 4.25 13.5 GB 15.6 GB Won't fit 5.00
MXFP4 4.25 13.5 GB 15.6 GB Won't fit 5.00
Q4_K_M 4.85 13.4 GB 15.5 GB Won't fit 5.03
Q4_K_S 4.58 12.7 GB 14.8 GB Won't fit 5.52
Q4_0 4.55 12.6 GB 14.7 GB Won't fit 5.55
IQ4_XS 4.25 11.9 GB 13.9 GB Won't fit 6.16
Q3_K_M 3.91 11.0 GB 13.1 GB Won't fit 7.26
IQ3_M 3.70 10.4 GB 12.5 GB Won't fit 8.01
IQ3_XXS 3.06 8.8 GB 10.9 GB Won't fit 12.6
Q2_K 2.63 7.7 GB 9.8 GB Won't fit 512 19.6
IQ2_XXS 2.06 6.2 GB 8.3 GB Fits, but tight 10K 68.7
IQ1_M 1.75 5.4 GB 7.5 GB Runs comfortably 15K 77.6

Also worth checking